Notes

On September 7, 2026, a review of the SPY fund was done. This review looked at the fund's performance from 2016 to September 4, 2026, using 15-minute updates. The data was collected with a system to prevent overload and with limits on retries. The numbers are observations and not investment advice or predictions. They are based on raw data and daily adjustments to wealth, using levels from the completed trading day on September 4. The SPY fund's details on September 4, 2026, were as follows: the symbol is SPY, the date is 2026-09-04, the close price is 770.19, the simple moving average over 200 days is 712.1529, the relative strength index over 2 days is 56.24216820150644, the previous relative strength index over 2 days were 52.00126399600207, 85.09673780322036, and 56.24216820150644, the average true range over 20 days is 5.236859999999979, the average daily range over 20 days is 4.4095599999999875, the volume in US dollars is 28041884838.015, and the whole shares are 0. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, for a 2021-2025 chronological holdout, which is a period of time used to test a strategy, with a main 65 and 10 basis points, which is a unit of measurement for interest rates. This used NumPy, a math software, with 20,000 circular 3-calendar-month blocks, where inactive months were kept, and a type of regression called HAC3 via sandwich covariance, which is a statistical method. A simulation was done with a hypothetical 25 percent account allocation, starting with 850.99, where resampled frequencies were not used to make forecasts, and monthly and realized marks did not include open-trade excursions, which are trades that have not been closed, and fixed overhead, which are costs that do not change. The simulation resampled realized cash flows over 24 days, but did not make a forecast for new entries. The actual size was 0. The results included 4 target counts, 19 stop counts, 4 RSI counts, and 0 timeout counts, with a target rate of 14.81 percent, and a target Wilson 95 percent of 5.92 and 32.48 percent. There was 1 same-day event, a stop overshoot median R of 0.049, a stop overshoot max R of 0.281, and 14 extended exits. A news scan found the data line: {'target_count': 4,'stop_count': 19, 'rsi_count': 4, 'timeout_count': 0, 'target_rate_pct': 14.814814814814813, 'target_Wilson95_pct': [5.91583185964489, 32.479063346829484],'same_day': 1,'stop_overshoot_median_R': 0.048566921876206606,'stop_overshoot_max_R': 0.28112018879816886, 'extended_exits': 14} https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following simulation details: Calculated on September 7, 2026, this is a 2021-2025 chronological holdout, with a main 65/10 basis point setup. The simulation used NumPy to create 20,000 circular 3-calendar-month blocks, keeping inactive months, and used HAC3 regression via sandwich covariance, which is a statistical method to estimate the relationship between variables. The simulation assumed a hypothetical 25% account allocation, starting with $850.99, and resampled frequencies were not calibrated forecasts. Monthly and realized marks did not include open-trade excursions and fixed overhead. This 24-day simulation resampled realized cash flows, and was not a fresh-entry forecast. The actual size was 0. The simulation results included 20,000 paths, 60 months, and a seed of 90706, with a hypothetical fraction of 0.25. The results also showed profit and loss quantiles, a loss probability of 0.98835, and a 95th percentile drawdown of 3.8067%. Additionally, the 24-day profit and loss quantiles were reported, along with the probability of a 24-day gain and loss. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. The main calculation used a 65/10 basis point method. A type of statistical analysis called HAC3 regression was used, which involves a sandwich covariance method. This analysis used 20,000 blocks of three-month data, and months where no trades were made were still included. A simulation was run, which assumed a 25% allocation of a hypothetical account with a starting value of $850.99. The simulation used realized marks, which are the actual values of the trades, and did not include open-trade excursions or fixed overhead costs. The simulation was run 24 times, and each time it used a different sample of realized cash flows. The results of the simulation are as follows: there were four regimes, or periods of time with different market conditions, which were high volatility, low volatility, RSI below 10, and RSI at least 10. The results for each regime were: high volatility had 19 occurrences, with a mean percentage change of -0.2525% and a win percentage of 31.57%, low volatility had 8 occurrences, with a mean percentage change of -0.4817% and a win percentage of 12.5%, RSI below 10 had 18 occurrences, with a mean percentage change of -0.1714% and a win percentage of 33.33%, and RSI at least 10 had 9 occurrences, with a mean percentage change of -0.6186% and a win percentage of 11.11%. The results by year were: 2021 had 8 occurrences, with a return percentage of -0.6383% and a mean percentage of -0.0736%, 2023 had 1 occurrence, with a return percentage of -1.6895% and a mean percentage of -1.6895%, 2024 had 12 occurrences, with a return percentage of -4.1656% and a mean percentage of -0.3498%, and 2025 had 6 occurrences, with a return percentage of -2.1674% and a mean percentage of -0.3626%. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. This analysis used a method called HAC3 regression, which is a statistical technique, via sandwich covariance, which is a way to calculate variance. The simulation assumed a 25% allocation of a hypothetical account with a starting value of $850.99. The results are based on 24-day simulation resamples of realized cash flows, not a forecast for new investments. A news scan found the following data: a dictionary with several values, including the number of observations, which is 27, the mean percentage, which is -0.3204%, and the mean confidence interval percentage, which ranges from -0.6802% to -0.0449%. The data also includes a positive bootstrap fraction of 0.0119, a volatility threshold of 0.1068, and coefficients for a statistical model called OLS, which stands for Ordinary Least Squares, a method for estimating relationships between variables. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. This was a chronological holdout, which is a way of testing a model by leaving out some data, with a main focus on 65 and 10 basis points, which are units of measurement for interest rates. The calculation used a type of statistical analysis called HAC3 regression via sandwich covariance, which is a method for estimating the relationship between variables. The analysis was done using 20,000 blocks of three calendar months, with inactive months included. A simulation was run, which is a computer model that mimics real-world scenarios, with a hypothetical 25% account allocation and a starting value of 850.99. The simulation used resampled frequencies, which means that the data was reorganized to create new samples, but these were not used to make forecasts. The simulation also used monthly and realized marks, which are values that are used to track the performance of an investment, and omitted open-trade excursions, which are trades that are still open, and fixed overhead, which are costs that remain the same even if the investment changes. The simulation ran for 24 days and resampled realized cash flows, which are the actual flows of money into or out of an investment, but this was not a forecast of what would happen if new investments were made. The actual size of the investment was 0. The results of the simulation included various counts and rates, such as a target count of 0, a stop count of 0, and an RSI count of 20, which is a measure of the speed and change of price movements. The target rate percentage was 0, and the target Wilson 95 percentage was between -1.3877787807814457e-15 and 16.113012549493323, which are statistical measures of the uncertainty of the results. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following simulation details: Calculated on September 7, 2026, this is a 2021-2025 chronological holdout, with a main 65 and 10 basis points. It uses NumPy with 20,000 circular 3-calendar-month blocks, keeping inactive months, and HAC3 regression via sandwich covariance. The simulation assumes a hypothetical 25% account allocation, starting at $850.99. It resamples realized cash flows over 24 days, without forecasting new entries, and omits open-trade excursions and fixed overhead. The simulation results are as follows: it ran 20,000 paths over 60 months, with a seed of 90706, and a hypothetical fraction of 0.25. The results include profit and loss quantiles, a loss probability of 0.91215, and a 95th percentile drawdown of 5.347101675364495%. The 24-day simulation results show specific profit and loss quantiles, and probabilities of gain and loss. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. The main calculation used a 65/10 basis point method. The NumPy software was used to create 20,000 circular blocks of three calendar months, and months that were inactive were kept. A HAC3 regression was done using sandwich covariance. A simulation was run with a hypothetical 25% account allocation, starting with $850.99. The simulation used resampled frequencies, but these were not used to make forecasts. The monthly and realized marks did not include open-trade excursions and fixed overhead. The simulation was run for 24 days and resampled realized cash flows, but it was not a forecast for new entries. The actual size was 0. The results were divided into regimes, which are groups of similar market conditions. The regimes were: high volatility with 16 instances, a mean percentage change of -0.5231457097038483, and a win percentage of 56.25; low volatility with 4 instances, a mean percentage change of -0.13013390788667312, and a win percentage of 50; RSI below 10 with 9 instances, a mean percentage change of -0.750386241224056, and a win percentage of 55.55555555555556; and RSI at least 10 with 11 instances, a mean percentage change of -0.1943082559810693, and a win percentage of 54.54545454545454. The results were also divided by year: 2021 had 7 instances, a return percentage of 0.004544791709082396, and a mean percentage of 0.006477822445391126; 2023 had 1 instance, a return percentage of -0.28418962055751784, and a mean percentage of -0.28418962055751784; 2024 had 7 instances, a return percentage of -8.72410952510666, and a mean percentage of -1.2678289504404083; and 2025 had 5 instances, a return percentage of 0.20925858596791969, and a mean percentage of 0.04455610594287451. RSI stands for Relative Strength Index, which is a measure of market conditions. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. This used a method called NumPy to look at blocks of three calendar months, keeping inactive months. A type of analysis called HAC3 regression was used, which involves something called sandwich covariance, a way to measure the variation of data. A simulation was run, assuming 25% of an account was allocated to this, starting with $850.99. This simulation looked at how things might have gone in the past, but did not try to predict the future. It used real data from the past, ignoring any open trades and fixed costs. The simulation was run 24 times, using real cash flow data, not trying to predict what would happen if someone entered the market at that point. The actual results were: the average percentage change was -0.4445%, with a range of -1.4272% to 0.1655%. Only 8.91% of the simulations had positive results. The volatility threshold was 0.1068. A statistical method called OLS was used to analyze the data, which found coefficients for constant, volatility, and something called RSI, a measure of how much something has changed recently. The results of this analysis were: constant was -0.0189, volatility was 0.0224, and RSI was 0.0009. Another statistical measure, OLS HAC t, found values of -1.224 for constant, 0.249 for volatility, and 2.143 for RSI. The R-squared value, which measures how well the data fits the model, was 0.1491. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline 'EEM QR290706 swing holdout exits'. This refers to a trading strategy for the EEM fund. Calculated on September 7, 2026, the strategy used data from 2021 to 2025. The main strategy used a 65/10 basis point approach, which means it considered a 65 basis point gain and a 10 basis point loss. The strategy used a type of math called NumPy to analyze 20,000 blocks of three calendar months, keeping inactive months. It also used a statistical method called HAC3 regression via sandwich covariance, which helps to understand the relationships between different variables. The simulation assumed a 25% allocation of a hypothetical account with a starting value of $850.99. The simulation did not try to predict the future, but rather looked at what would have happened in the past. It also did not consider the costs of opening and closing trades. The simulation ran for 24 days and looked at the actual cash flows, not trying to predict what would happen if new trades were made. The results included the following statistics: there were 5 target counts, 17 stop counts, 1 RSI count, and 0 timeout counts. The target rate was 21.74%. The target Wilson 95% range was between 9.66% and 41.90%. There was 1 same-day trade. The median stop overshoot was 0.049 and the maximum stop overshoot was 0.918. There were 8 extended exits. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following simulation details: Calculated on September 7, 2026, this is a 2021-2025 chronological holdout, with a main 65/10 basis point setup. The simulation used NumPy to create 20,000 circular 3-calendar-month blocks, keeping inactive months, and a HAC3 regression via sandwich covariance. The simulation assumed a hypothetical 25% account allocation, starting at $850.99. It resampled frequencies, which are not calibrated forecasts, and used monthly and realized marks, omitting open-trade excursions and fixed overhead. The 24-day simulation resampled realized cash flows, and is not a fresh-entry forecast. The actual size is 0. The simulation results include 20,000 paths, 60 months, and a seed of 90706, with a hypothetical fraction of 0.25. The results also show profit and loss quantiles, a loss probability of 0.8572, and a 95th percentile drawdown of 3.5512%. Additionally, the 24-day profit and loss quantiles and probabilities of gain and loss are provided. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. This analysis used a method called NumPy to look at 20,000 blocks of three calendar months, keeping months when nothing happened. A statistical method called HAC3 regression was used, which involves something called sandwich covariance, a way to estimate variability. The simulation assumed putting 25% of a hypothetical account into this, starting with $850.99. It did not try to predict the future, but instead looked at what actually happened, ignoring open trades and fixed costs. The simulation re-examined actual cash flows over 24 days, not trying to forecast new entries. The actual size of the investment was zero. The results were broken down into regimes, which are like categories. The regimes were: high volatility with 4 instances, averaging a -0.1226% return, and winning 25% of the time; low volatility with 19 instances, averaging a -0.2523% return, and winning 31.57% of the time; RSI, or Relative Strength Index, below 10 with 17 instances, averaging a 0.0033% return, and winning 41.18% of the time; and RSI at least 10 with 6 instances, averaging a -0.8901% return, and winning 0% of the time. The results were also broken down by year: 2021 had 4 instances, a 0.7097% return, and averaged 0.1930% per instance; 2023 had 5 instances, a -2.5976% return, and averaged -0.5239% per instance; 2024 had 7 instances, a -4.3923% return, and averaged -0.6364% per instance; and 2025 had 7 instances, a 0.9835% return, and averaged 0.1454% per instance. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, looking at data from 2021 to 2025. This used a method called NumPy to analyze 20,000 blocks of three calendar months, keeping months that were inactive. A type of regression called HAC3 was used, which involves something called sandwich covariance, a way to estimate the variance of a regression. A simulation was run, assuming 25% of an account was allocated to this, starting with $850.99. This simulation did not try to predict the future, but rather looked at what would have happened if this strategy had been used in the past. It looked at the actual cash flows, not trying to predict what would happen if someone were to start this strategy now. The simulation was run 24 times, looking at the results of using this strategy in different time periods. The results were: there were 23 data points, the average percentage change was -0.22975146312720546%, the average confidence interval was between -0.665411822482381% and 0.20920510443481755%, 14.465% of the simulations had a positive result, the threshold for volatility was 0.16887184166767816%, and a type of analysis called OLS found coefficients for constant, volatility, and something called rsi, which stands for relative strength index, a measure of how strong a stock is. The OLS analysis also found t-values and an R-squared value of 0.2943475794528686. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. This was a test of a trading strategy, using a method called HAC3 regression, which is a way to analyze data. The test used a computer program called NumPy to look at blocks of three calendar months. The test assumed that 25% of a hypothetical account would be invested, starting with $850.99. The results are based on what actually happened, not on a forecast of what might happen in the future. A news scan found the following data line: a 24-day simulation showed the results of the trading strategy, but this was not a forecast of what would happen if new trades were made. The results included the following numbers: target count was 0, stop count was 0, rsi count was 35, timeout count was 0, target rate percentage was 0, and target Wilson 95 percentage was between 0 and 9.89%. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following simulation details: Calculated on September 7, 2026, this is a 2021-2025 chronological holdout, with a main 65/10 basis point setup. The simulation used NumPy to create 20,000 circular 3-calendar-month blocks, keeping inactive months, and a HAC3 regression via sandwich covariance, which is a statistical method to estimate the relationship between variables. The simulation assumed a hypothetical 25% account allocation, starting with $850.99. It resampled frequencies, but did not calibrate forecasts, and omitted open-trade excursions and fixed overhead from monthly and realized marks. The 24-day simulation resampled realized cash flows, rather than making a fresh-entry forecast. The actual size is 0. The simulation results included 20,000 paths, 60 months, and a seed of 90706, with a hypothetical fraction of 0.25. The results also showed profit and loss quantiles, a loss probability of 0.47345, and a 95th percentile drawdown of 6.35024605440212%. Additionally, the results included 24-day profit and loss quantiles, and probabilities of gain and loss. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. This analysis used a method called NumPy to look at 20,000 blocks of three calendar months, keeping the months when nothing happened. It also used a statistical method called HAC3 regression via sandwich covariance, which is a way to account for variations in the data. The simulation assumed putting 25% of a hypothetical account into the investment, starting with $850.99. The results are based on what actually happened, not on predictions of what might happen in the future. The simulation looked at the money made or lost over 24 days, but did not try to predict what would happen if new investments were made. The actual size of the investment was zero. The results were divided into groups, or regimes, based on certain characteristics. The regimes were: high volatility, low volatility, and two based on a measure called the Relative Strength Index, or RSI, which is a way to measure how quickly prices are changing. The results for each regime were: high volatility with 12 events, an average loss of 0.79%, and 50% of the events making money; low volatility with 23 events, an average gain of 0.44%, and 65.22% of the events making money; RSI below 10 with 16 events, an average gain of 0.58%, and 56.25% of the events making money; and RSI at least 10 with 19 events, an average loss of 0.45%, and 63.16% of the events making money. The results were also looked at by year: 2021 had 6 events, a total return of 2.77%, and an average gain of 0.46%; 2023 had 9 events, a total loss of 1.94%, and an average loss of 0.18%; 2024 had 10 events, a total loss of 4.19%, and an average loss of 0.41%; and 2025 had 10 events, a total gain of 3.54%, and an average gain of 0.37%. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. This was a chronological holdout, which is a method of testing a model, with a main allocation of 65 and a 10 basis point, or 0.1%, threshold. The calculation used 20,000 circular blocks of three calendar months, keeping inactive months, and a type of regression called HAC3, which uses sandwich covariance, a way to estimate the variance of a model. A simulation was run with a hypothetical 25% account allocation starting at $850.99. The simulation resampled realized cash flows over 24 days, but did not forecast new entries. The results were: the number of observations was 35, the mean percentage was 0.0199%, the mean confidence interval percentage was between -0.85% and 0.615%, the fraction of positive bootstrap samples was 0.5319, the volatility threshold was 0.1689, and the ordinary least squares, or OLS, coefficients were calculated, with a constant term of 0.0102, a volatility term of -0.0821, and a relative strength index, or RSI, term of 0.000189. The OLS HAC t-values were also calculated, with a constant term of 0.9017, a volatility term of -1.0345, and an RSI term of 0.3913. The OLS R-squared value was 0.0299. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, for a chronological holdout from 2021 to 2025, with a main 65/10 basis point. This used NumPy with 20,000 circular three-calendar-month blocks, keeping inactive months. A HAC3 regression was done via sandwich covariance. A simulation was run with a hypothetical 25 percent account allocation starting at 850.99. This simulation did not make new forecasts, but instead used realized marks, omitting open trade excursions and fixed overhead. The simulation resampled realized cash flows over 24 days, not making a fresh entry forecast. The actual size was 0. The results included 6 target counts, 23 stop counts, 6 RSI counts, and 0 timeout counts. The target rate was 17.14 percent. The target Wilson 95 percent was between 8.10 percent and 32.68 percent. There were 8 same day events. The median stop overshoot was 0.16 and the maximum stop overshoot was 1.83. There were 17 extended exits. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following simulation details: Calculated on September 7, 2026, this is a 2021-2025 chronological holdout, with a main 65/10 basis point setup. The simulation used NumPy to create 20,000 circular 3-calendar-month blocks, keeping inactive months, and used HAC3 regression via sandwich covariance, which is a statistical method to estimate the relationship between variables. The simulation assumed a hypothetical 25% account allocation, starting with 850.99, and resampled frequencies without calibrated forecasts, omitting open-trade excursions and fixed overhead. This 24-day simulation resampled realized cash flows, not a fresh-entry forecast. The actual size is 0. The simulation results include 20,000 paths, 60 months, and a seed of 90706, with a hypothetical fraction of 0.25. The results also show profit and loss quantiles, a loss probability of 0.99375, a 95th percentile drawdown of 6.379963854917175 percent, and 24-day profit and loss quantiles. Additionally, the results show a probability of a 24-day gain of 0 and a probability of a 24-day loss of 0.19275. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. The main calculation used a 65/10 basis point method. The calculation used 20,000 blocks of three calendar months, and months with no activity were kept. A type of analysis called HAC3 regression was used, which involves sandwich covariance, a way to measure the variation of data. A simulation was run with a hypothetical 25% account allocation starting at $850.99. The simulation used resampled frequencies, which are not the same as forecasts, and monthly marks that did not include open trade excursions or fixed overhead costs. The simulation ran for 24 days and resampled realized cash flows, which is not the same as a forecast for new entries. The actual size was zero. The results were divided into regimes, which are groups of data with similar characteristics. The regimes were: high volatility with 16 instances, a mean percentage change of -0.3536, and a win percentage of 43.75; low volatility with 19 instances, a mean percentage change of -0.4668, and a win percentage of 21.05; RSI below 10 with 23 instances, a mean percentage change of -0.3158, and a win percentage of 34.78; and RSI at least 10 with 12 instances, a mean percentage change of -0.6052, and a win percentage of 25. The results were also divided by year: 2021 had 9 instances, a return percentage of -4.13, and a mean percentage of -0.4612; 2022 had 1 instance, a return percentage of 0.65, and a mean percentage of 0.6479; 2023 had 7 instances, a return percentage of -2.13, and a mean percentage of -0.3019; 2024 had 8 instances, a return percentage of -1.67, and a mean percentage of -0.2044; and 2025 had 10 instances, a return percentage of -7.10, and a mean percentage of -0.7275. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data: Calculated on September 7, 2026, this analysis covers the period from 2021 to 2025. It uses a 65/10bp main holdout and NumPy to create 20,000 circular 3-calendar-month blocks, keeping inactive months. The analysis uses HAC3 regression via sandwich covariance. A simulation was run with a hypothetical 25% account allocation starting at $850.99. The simulation resamples realized cash flows over 24 days, but does not forecast fresh entries. The results are as follows: the mean percentage is -0.4150340515971169, the mean confidence interval percentage is between -0.7320018622318439 and -0.09901803554388251, the positive bootstrap fraction is 0.0064, and the volatility threshold is 0.11832438027994344. The ordinary least squares coefficients are -0.0037781664635392143 for the constant, 0.005533750541884458 for volatility, and -0.00010073221152994208 for the relative strength index. The ordinary least squares HAC t-values are -0.33640951793310647 for the constant, 0.058458526366506065 for volatility, and -0.6314986002767219 for the relative strength index. The ordinary least squares R2 is 0.0049254611746821775. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. This was a chronological holdout, which is a way of testing a model, with a main allocation of 65 and a 10 basis point, or 0.1 percent, adjustment. The calculation used a type of statistical analysis called HAC3 regression, which involves a sandwich covariance, a way of estimating the variability of the data. The simulation assumed a 25 percent allocation of a hypothetical account with a starting value of 850.99. The results were based on resampled frequencies, which are repeated samples of the data, and did not include forecasts or the value of open trades and fixed costs. The simulation ran for 24 days and resampled the actual cash flows, rather than predicting new entries. The actual size of the allocation was 0. The results included the following statistics: target count was 0, stop count was 0, rsi count, which refers to the Relative Strength Index, a measure of stock price movement, was 35, timeout count was 0, target rate percent was 0, and target Wilson 95 percent, a statistical measure of confidence, was between 0 and 9.89 percent. Other statistics, including same day, stop overshoot median R, stop overshoot max R, and extended exits, were either 0 or null, meaning they had no value. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline 'EFA QR290706 nextopen holdout simulation'. This is a simulation result. Calculated on September 7, 2026, it uses data from 2021 to 2025. The main calculation uses a 65/10 basis point method. The simulation uses a type of math called NumPy to look at 20,000 blocks of three calendar months, keeping months when nothing happened. It uses a statistical method called HAC3 regression via sandwich covariance, which is a way to calculate relationships between things. The simulation assumes putting 25% of a hypothetical account into this, starting with 850.99. It looks at what actually happened, not trying to predict the future, and ignores open trades and fixed costs. The simulation replays what happened over 24 days, not trying to predict new entries. The actual size is zero. The simulation has many details, including 20,000 paths, 60 months, and a seed of 90706. It also has a hypothetical fraction of 0.25, and many calculated numbers, such as a loss probability of 0.41695 and a drawdown p95 percentage of 6.794152938183459. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. This was a chronological holdout, meaning some data was set aside, with a main allocation of 65 and a 10 basis point threshold. The calculation used 20,000 circular blocks of three calendar months, keeping inactive months, and a type of regression called HAC3 via sandwich covariance. A simulation was run with a hypothetical 25% account allocation starting at $850.99. The simulation resampled frequencies, but did not make calibrated forecasts, and omitted open-trade excursions and fixed overhead from monthly and realized marks. The simulation ran for 24 days, resampling realized cash flows, and was not a forecast for new entries. The actual size was 0. The results were divided into regimes, which are categories of market conditions. The regimes were: high volatility, low volatility, RSI below 10, and RSI at least 10. The results for each regime were: high volatility with 18 occurrences, a mean percentage change of -0.1836, and a win percentage of 61.11; low volatility with 17 occurrences, a mean percentage change of 0.3494, and a win percentage of 64.71; RSI below 10 with 21 occurrences, a mean percentage change of 0.4680, and a win percentage of 76.19; and RSI at least 10 with 14 occurrences, a mean percentage change of -0.5140, and a win percentage of 42.86. The results were also divided by year: 2021 with 11 occurrences, a return percentage of 9.11, and a mean percentage of 0.8047; 2022 with 1 occurrence, a return percentage of 0.6479, and a mean percentage of 0.6479; 2023 with 7 occurrences, a return percentage of 1.67, and a mean percentage of 0.2485; 2024 with 8 occurrences, a return percentage of -6.60, and a mean percentage of -0.8362; and 2025 with 8 occurrences, a return percentage of -2.32, and a mean percentage of -0.2395. https://runway.adityaperswal.workers.dev/data/stock-bars

A calculation was made on September 7, 2026, using data from 2021 to 2025. This analysis used a method called HAC3 regression, which is a statistical technique that accounts for uncertainty, via sandwich covariance. The analysis was done in blocks of three calendar months, and months where no activity happened were still included. A simulation was run with a hypothetical 25% allocation of an account, starting with $850.99. This simulation did not try to predict the future, but instead looked at what would have happened if the same actions were taken in the past. The results are based on 24 days of simulated trading, and do not include the results of opening new trades or fixed costs. The results of the simulation are as follows: there were 35 data points, with an average return of 0.07524363740623422%, and a range of possible average returns between -0.8456935251601032% and 0.7495949283961985%. A news scan found the headline was not applicable here, instead the results show that 58.81% of the simulations had positive results. The volatility threshold was 0.11832438027994344. A statistical model called OLS was used to analyze the results, and it found that the constant term was 0.017367187777731125, the coefficient for volatility was -0.09423304753595095, and the coefficient for the relative strength index, which is a measure of how quickly prices are changing, was -0.0004141020031927928. The t-values for these coefficients were 0.6558313820458799 for the constant term, -0.4409216389128227 for volatility, and -1.4511014491652028 for the relative strength index. The R-squared value, which measures how well the model fits the data, was 0.03934245843859474. https://runway.adityaperswal.workers.dev/data/stock-bars